Research on the technology of image generation for imaging guidance simulation based on DSP

The development of a new generation of precision-guided weapon systems has put forward new requirements for optical imaging guidance simulation target simulators, which requires the target/background imaging simulation system to provide real-time dynamic scenes. The physical simulation of missiles requires not only expensive missiles or models, but also a large shooting range, and a large number of personnel to perform a series of tasks such as observation, recording and control; and once launched, the process is irreversible, which is a very expensive, time-consuming and laborious task. Among them, DSP-based image generation technology is a new and key technology that can reduce R&D costs, reduce risks, and provide a strong guarantee for the advancement of system technology. Here we mainly give the dynamic scenes generated by image generation technology, and simulate the changes of dynamic scenes when the missile is working in pitch and heading directions.

1 Design of Imaging Guidance Simulation System

1.1 System composition

In order to realize the real-time display of the simulated image, the selection of DSP is the key to the design of the DSP video generation system. For this reason, the TMS320C6418 DSP is selected here. The system principle block diagram of the image generation simulation system with TMS320C6418 as the core is shown in Figure 1.

1.2 System Working Principle

When the system is initialized, the material image is sent to the memory SDRAM through DSP to wait for DSP processing, and then the processed image data is sent to the FPGA RAM through the DSP's EDMA channel under the control of the horizontal synchronization and field synchronization signals generated by the FPGA. Under the control of the composite synchronization signal and composite blanking signal generated by the FPGA, the image data is converted by D/A and synthesized into a standard analog video signal.

1.3 DSP-based system hardware characteristics

The DSP processing system processes image data and is the core of the entire system. It consists of DSP device TMS320C6418, image memory SDRAM, and program memory Flash.

TMS320C6418 is connected to SDRAM, Flash and FPGA through external memory interface (EMIF) to form a typical embedded DSP application system. TMS320C6418 is a new TMS320C64xx series launched by TI. Its main features are: 512K bytes of RAM on chip; operation speed up to 600MHz; support for multiple external memory interfaces, including SRAM, SDRAM, RAOM, FIF-0; enhanced EDMA controller with 64 EDMA channels.

In real-time image digital processors, using SDRAM to store image data has two advantages: first, high speed, with access time up to 8 ns; second, large single-chip capacity, which helps reduce the size of the image processing system.

For DSP processing system, it is necessary to expand the off-chip program memory, and Flash memory is used here. The RAM (L2) capacity of TMS320C6418 reaches 512 KB, and the application program of DSP system is usually not very large. Therefore, after the system is powered on, the used program is loaded into L2, which can improve the access speed of the program.

1.4 DSP-based system imaging guidance simulation image generation software

1) The DSP initialization program completes the initialization work after power-on reset, mainly including the configuration and initialization of external storage devices SDRAM and Flash and moving the material image data into SDRA-M.

2) DSP system image data transmission program There is a large amount of data transmission in image generation, including: target and background data transmission to realize image data fusion processing; data transmission between SDRAM and DSP internal RAM to realize target extraction processing; data transmission between SDRAM and FPGA RAM to realize real-time display of video images.

3) The image processing program selects the appropriate image processing algorithm according to the characteristics of the image, extracts the target, and provides the control end with the miss distance to achieve closed-loop tracking of the image.

4) The guidance control program is based on the input conditions of external input, namely tracking, scanning and other commands, so that it can simulate the dynamic changes of the real scene when the seeker is working.

2 Key technologies for image generation of imaging guidance simulation images

2.1 Modeling of background and objectives

The target refers to the object of military weapons combat, while the background is all space materials except the target. The target and the background are relative, and the same object can be a target sometimes and a background sometimes. For imaging guidance, according to the different heights or properties of the geographical locations of the target and the background, the target and the background can be divided into three categories: aerial targets and sky background, sea surface targets and ocean background, and ground targets and ground background.

The background image can be generated using a regular rectangular raster image, such as a *.bmp bitmap image with sky and ground. In some specific scenes, irregular target raster images such as flowers, grass, trees, cars, houses, etc. are also required. For example, if you want to generate a big tree in an existing scene, how can you see the scenery behind through the gaps between leaves and branches? There is no special function in DSP to handle this. However, it can be achieved by the following methods:

1) First, use 3DMAX software to generate the required flowers, grass, cars, etc., and generate files in *.BMP bitmap file format. Use PHOTOSHOP software to adjust the contrast and brightness of the image of interest. Make the bitmap image into a rectangle to facilitate DSP algorithm processing.

2) Fusion of background and target: After determining the position of the target, the background and target need to be merged. However, due to the brightness difference between the background and the target, if they are simply overlapped, there will be obvious gaps. In order to achieve seamless fusion, the edges of the overlapping areas of the background and the target are synthesized into new edges according to a certain weight value.

Where Inew is the synthesized new edge image, I1 and I2 are the edge images of the overlapping area between the target and the background, and w is the weighting coefficient.

Assuming the width of the overlapping edge area is d, then the stitched image, I1 transitions to I2, w changes linearly from 0 to 1, with an increment of 1/d. This algorithm achieves a uniform transition between the target and background edges, achieving a seamless fusion effect.

3) Filter the generated image.

2.2 Generation of material image data

Various background images and target images are stored in the memory, which requires converting the acquired materials into digital images and then burning these materials into the memory through the DSP through a burning program.

1) Use software to convert the material into a data format that can be recognized by TI CCStudio debugging software. The specific format is shown in Figure 2.

2) Use CCStudio's File->Data->Load menu to load the material data into the corresponding address space.

3) The material data is solidified into the corresponding address space of the memory through the written burning program.

3 Real-time simulation of guidance loop images

To realistically simulate the combat environment, the various movement laws of the target must be considered. There are three ways to describe the movement of the target: 1) the target moves, and the seeker frame does not move; 2) the target does not move, and the seeker frame moves; 3) the target and the seeker frame move at the same time.

In the missile simulation loop of artificial target capture, the target is usually fixed, and the missile is controlled by manual operation to search and track the target. Therefore, this paper adopts the second method (i.e. the target is stationary, and the seeker frame moves) to perform image simulation. The simulation process is shown in Figure 3.

After the simulation image is generated, it is sent to the image processing module. The image processing module uses image processing and recognition technology to complete the detection, recognition and tracking of the target, and gives the corresponding error signal according to the target position; the control module performs the azimuth and pitch scanning of the seeker according to the corresponding control command, and immediately switches to the tracking state once the target is found, and calculates the position and attitude parameters of the seeker at the same time; the adjustment result of the seeker is sent back to the simulation field of view image generation module as feedback, and the module calculates and generates the next image according to the feedback signal. The actual effect of the simulation is shown in Figure 4.

4 Conclusion

This paper introduces the imaging guidance image generation simulation system based on DSP processor, and applies it in a seeker simulation system, which truly simulates the whole process of the seeker searching and capturing the target. This simulation system can judge the target recognition and guidance capabilities of the seeker, and thus improve the target recognition system, thereby improving the precision guidance capability. With the development of precision-guided weapon systems, DSP-based imaging guidance simulation will surely make greater and greater progress and play an increasingly important role.